Mutant prevention and minimum inhibitory concentration drug values for enrofloxacin, ceftiofur, florfenicol, tilmicosin and tulathromycin tested against swine pathogens Actinobacillus pleuropneumoniae, Pasteurella multocida and Streptococcus suis.

Actinobacillus pleuropneumoniae, Pasteurella multocida and Streptococcus suis are prevalent bacterial causes of swine infections. Morbidity, mortality and positively impacting the financial burden of infection occurs with appropriate antimicrobial therapy. Increasing antimicrobial resistance complicates drug therapy and resistance prevention is now a necessity to optimize therapy and prolong drug life. Mutant bacterial cells are said to arise spontaneously in bacterial densities of 107-109 or greater colony forming units/ml. Antibiotic drug concentration inhibiting growth of the least susceptible cell in these high density populations has been termed the mutant prevention concentration (MPC). In this study MPC and minimum inhibitory concentration (MIC) values of ceftiofur, enrofloxacin, florfenicol, tilmicosin and tulathromycin were determined against the swine pathogens A. pleuropneumoniae, P.multocida and S. suis. The following MIC90/MPC90 values (mg/L) for 67 A. pleuropneumoniae and 73 P. multocida strains respectively were as follows: A. pleuropneumoniae 0.031/0.5, ≤0.016/0.5, 0.5/2, 4/32, 2/32; P. multocida 0.004/0.25, 0.016/0.125, 0.5/0.5, 8/16, 0.5/1. For 33 S. suis strains, MIC90 values (mg/L) respectively were as follows: 1, 0.25, 4, ≥8 and ≥8. A total of 16 S. suis strains with MIC values of 0.063-0.5 mg/L to ceftiofur and 0.25-0.5 mg/L to enrofloxacin were tested by MPC; MPC values respectively were 0.5 and 1 mg/L respectively. MPC concentrations provide a dosing target which may serve to reduce amplification of bacterial subpopulations with reduced antimicrobial susceptibility. Drug potency based on MIC90 values was ceftiofur > enrofloxacin >florfenicol = tulathromycin > tilmicosin; based on MPC90 values was enrofloxacin > ceftiofur > tulathromycin > florfenicol ≥ tilmicosin.


Introduction
Bacterial infectious diseases are complicated by antimicrobial resistance and global concerns with the clinical impact of resistance is redefining antimicrobial utilization [1,2]. In addition to clinical outcomes, microbiological measurements continue to contribute to novel data on drug use for treatment and bacterial eradication. Optimization of therapy needs to consider clinical outcomes and antimicrobial resistance prevention during therapy. Guideline documents for antimicrobial therapy of human infectious diseases give consideration of antimicrobial agents with a reduced likelihood for resistance selection to be important when making therapeutic choices [3,4]. Such an observation clearly impacts economic costs and adds substantially to treatment costs.
In pigs, respiratory disease is amongst the most important health concerns for swine producers. Swine respiratory disease has been previously recognized as the main pathogen-identified cause of swine mortality accounting for deaths in~44% of nursing pigs and~61% of grown finished pigs [5]. Porcine respiratory disease complex (PRDC) is a multifactorial clinical entity describing pneumonia in pigs where multiple etiologies contribute to the pathogenesis leading to clinical disease [5]. This complex etiology and pathogenesis may include one or more viruses, Mycoplasma hyopneumoniae, opportunistic and pathogenic bacteria. P. multocida is an important pig pathogen and is carried by a large number of animals and transmission is mostly by aerosols [6,7]. Following invasion, P. multocida multiplies quickly, liberates toxins and causes necrotic lesions in lung tissue. A. pleuropneumoniae is highly contagious and causes an acute or chronic fibro-haemorrhagic necrotising pneumonia [8]. Ceftiofur (beta-lactam), enrofloxacin (fluoroquinolone), florfenicol (phenol), tilmicosin (macrolide) and tulathromycin (triamalide) are commonly used for swine infections.
Antimicrobial susceptibility or resistance is determined in vitro by measuring the minimum inhibitory concentration (MIC) utilizes a bacterial inoculum of 10 5 colony forming units per millilitre (cfu/ml) [9]. Previous publications for human infectious diseases reported substantially higher bacterial densities (i.e. 10 7 −10 9 cfu/ml or higher) during infections such as meningitis, pneumonia and from protected brush specimens from patients with an acute bacterial exacerbation of their chronic lung disease [10][11][12][13]. Additionally, McVey and Kusak studied lung, tonsil and trachea tissues from calves with bronchopneumonia and reported 12% of samples had >10 8 cfu/g and 50% of specimens hads >10 5 cfu/g with Mannheimia haemolytica being the most common organism recovered [14]. Given the substantially higher bacterial densities in infection than tested in an MIC assay, it begs the question as to the true dynamics of bug/drug interactions when higher bacterial densities are encountered. The mutant prevention concentration (MPC) describes a drug concentration threshold or lowest drug concentration blocking growth of mutant bacterial sub-populations [15,16] that spontaneously arise in bacterial densities of 10 7 −10 9 cfu-densities seen with infection. Antibiotic drug concentrations insufficiently inhibiting mutant cell growth result in selective amplification of bacterial cells with reduced drug susceptibility [16,17]. In a study with fluoroquinolones and the human pathogen Streptococcus pneumoniae, differences were seen between fluoroquinolone compounds and macrolide compounds and MPC values [18,19]. Published MPC studies have been completed with human pathogens and fluoroquinolones, macrolides and many other drug classes [17,18,[20][21][22]. Fluoroquinolones used in veterinary medicine were previously tested by MPC against E. coli and Staphylococcus pseudintermedius and MPC measurements with M. haemolytica have been reported for ceftiofur, enrofloxacin, florfenicol, tilmicosin and tulathromycin showing differences between compounds in their ability to prevent mutant growth at clinically relevant concentrations [17,23,24].
Here we report on testing of swine clinical isolates of A. pleuropneumoniae, P. multocida and S. suis by MPC to ceftiofur, enrofloxacin, florfenicol, tilmicosin and tulathromycin to determine antimicrobial drug concentrations blocking the most resistant bacterial organisms in high density cultures. Observations reported here may inform thinking on antimicrobial use to affect clinical cure, minimize resistance selection during therapy and pharmacokinetic/ pharmacodynamic modelling.

Bacterial strains
Bacterial pathogens collected from swine in the U.S.A were used: A. pleuropneumoniae (n = 67), P. multocida (n = 73), S. suis (n = 59). These organisms were generously provided by Dr. Ching Ching Wu from the Indiana Animal Disease Diagnostic Laboratory, Purdue University, West Lafayette, Indiana. Bacterial strains were identified by Vitek II (bioMerieux, St. Laurent, QC), matrix assisted laser desorption ionization-time of flight (MALDI-TOF) (bioMerieux, St. Laurent, QC) and/or biochemical tests as summarized in the Manual of Clinical Microbiology [25]. Individual strains were stored at -70˚C in skim milk. For MIC testing, bacteria were thawed and subcultured two times on blood agar (tryptic soy agar containing 5% sheep red blood cells) (BA) plates with incubation for 18-24 hours at 35-37˚C in oxygen (O 2 ). Bacterial strains included in the study needed to be susceptible (where breakpoints exist) to the drugs tested by interpretative criteria as per the Clinical and Laboratory Standards Institute (CLSI) [9].

Antimicrobial compounds
Enrofloxacin was provided by Bayer Animal Health, Shawnee Mission, Kansas esd prepared as per manufacturer's instructions. Ceftiofur, florfenicol, tilmicosin and tulathromycin were purchased commercially through the Western College of Veterinary Medicine Pharmacy at the University of Saskatchewan and reconstituted based on manufacturer's directions. Fresh stock solutions or those prepared from frozen samples (-70˚C) were used. For quality control, the following American Type Culture Collection (ATCC) control strains were included in each susceptible assay to ensure performance of the susceptibility assays: Enterococcus faecalis ATCC 29212, Pseudomonas aeruginosa ATCC 27853, E. coli ATCC 25222, Staphylococcus aureus ATCC 29213. MIC values needed to be within acceptable ranges for each organism/ drug.

MIC measurements
MIC testing was based on the recommended CLSI procedure [9]. Briefly, Mueller-Hinton broth (MHB) containing two-fold concentration of drug was added to 96-well micro-dilution trays. A 0.5 McFarland density of A. pleuropneumoniae, P. multocida and S. suis was further diluted to 5 x 10 5 cfu/ml, added to the microdilution tray containing drug and incubated for 18-24 hours (35-37˚C) in O 2 . The MIC was interpreted as the lowest drug concentration inhibiting visible growth. The designation of MIC 50 and MIC 90 are determined by calculating the drug concentration inhibiting 50% or 90% of strains respectfully by starting from the lowest MIC or MPC values.

MPC testing
MPC testing was adapted from the method published for S. pneumoniae and previously reported for Mannheimia haemolytica [18,24]. Starter cultures for A. pleuropneumoniae and P. multocida were inoculated on 5 BA plates per isolate to produce confluent growth and then incubated at 35-37˚C for 18-24 hours in O 2 . Starter cultures for S. suis isolates were on 5 chocolate agar plates with incubation for 18-24 hours at 35-37˚C in O 2 following which the plate surfaces were swabbed to remove bacterial growth and transferred to 100 ml of brain heart infusion broth containing nicotinamide adenine dinucleotide (NAD) (A. pleuropneumoniae) or MHB (P. multocida) or Veterinary Fastidious Medium (MHB plus 3% laked horse blood) (Trek Diagnostic System, Cleveland, Ohio) (S. suis) and incubated as described. Following incubation, turbidity measurements verified cell densities of 3 x 10 8 cfu/ml. Centrifugation at 5000 x G for 3 minutes at 4˚C was used to concentrate bacteria following which the pellet was added to 3 ml of fresh medium. Drug containing agar plates (7 drug concentrations in doubling dilution) were inoculated with 200 ul (10 10 cfu) of bacterial suspension and incubated for 24 hours at 35-37˚C in O 2 and screened for growth. Plates were reincubated for an additional 24 hours and the final reading recorded. The MPC value was the lowest drug concentration blocking growth. Drug concentrations tested were ceftiofur 0.06 to 4 mg/L, enrofloxacin 0.004 to 2 mg/L, florfenicol 0.5 to 32 mg/L, tilmicosin 0.5 to 64 mg/L, tulathromycin 0.25 to 16 mg/L. The designation of MPC 50 and MPC 90 are by calculating the drug concentration inhibiting 50% or 90% of strains respectfully by starting from the lowest MIC or MPC values.

Results
MIC and MPC data for A. pleuropneumoniae strains and the 5 drugs is shown in Table 1 Table 2 summarizes MIC and MPC data for the P. multocida strain tested against the 5 drugs investigated. The MIC range , MIC 50 and MIC 90 values respectively for each agent were as follows: ceftiofur �0.016-0.031 mg/L, �0.016 mg/L, �0.016 mg/L; enrofloxacin �0.016 mg/L,   MPC investigations with various classes of antimicrobial agents have been reported [17,22,24,32] despite an earlier publication suggesting MPC measurements only apply to fluoroquinolones and not aminoglycosides, macrolides or beta-lactams [33]. For example, Metzler et al compared MPC values for azithromycin, clarithromycin and erythromycin against S. pneumoniae strains and showed clarithromycin was statistically less likely to select for organism with reduced susceptibility and azithromycin was statistically more likely [19].
A . From investigators with Streptococcus pneumoniae and moxifloxacin, resistance was selected at drug concentrations falling within the MSW and a more recent report found that time within the MSW was an appropriate prediction of bacterial resistance [35,38]. In a report from testing gatifloxacin and S. pneumoniae in a rabbit empyema model, dosing of gatifloxacin to remain within the MSW for �40% of the dose resulted in mutant subpopulation amplification [39].
We [24] and others [40,41]have previously commented on some of the characteristics of macrolide and macrolide like compounds (azalides, triamilides) in human and veterinary medicine. In particular, all such agents have low serum drug concentrations and lung concentrations vary when considering epithelial lining fluid drug concentrations versus alveolar macrophage drug concentrations. Our susceptibility results for tulathromycin appear consistent with results from previous measurements with M. haemolytica and more recently P. multocida [24,42]. In this study, the MIC 90 values for tilmicosin were 4 mg/L against the A. pleuropneumoniae and P. multocida strains as compared to 2 and 0.5 mg/L respectively for tulathromycin. MIC values in excess of 0.5 mg/L are above achievable or sustainable blood concentrations for tilmicosin and tulathromycin (www.zoetis.com).
For both tilmicosin and tulathromycin, MPC values were �8 mg/L for all strains of A. pleuropneumoniae. For tilmicosin MPC values were �2 mg/L for all P. multocida strains as compared to �0.5 mg/L for tulathromycin. For those compounds and considering the MSW and MIC 90 and MPC 90 values, the MSW for tilmicosin and A. pleuropneumoniae would range from 4 to 32 mg/L (8-fold difference) and for P. multocida from 4 to 16 mg/L (4-fold difference). By comparison, for tulathromycin and A. pleuropneumoniae the MSW would range from 2-32 mg/L (16-fold difference) and for P. multocida the MSW was narrow being the same or within a 2-fold difference. For individual strains, the MSW may be narrower if the fold difference between the measured MIC and MPC values were less.
In swine lungs, tulathromycin concentrations peak at approximately 3. [43] Given the achievable serum and pulmonary concentrations for tulathromycin, the susceptible breakpoints of 16 mg/L and 64 mg/L for P. multocida and A. pleuropneumoniae respectively is puzzling.
For the remaining drugs tested and considering the MSW, blood levels of ceftiofur exceed the MSW for 48-6 hours but we did not conduct testing to account for the high protein binding (>90%) associated with this compound. Others have shown the protein binding in excess of 60% elevates MIC values in vitro when protein is included in the susceptibility assays [46][47][48]. How this observation translates clinically is not fully understood. Enrofloxacin (7.5 mg/ kg) blood levels exceed the MSW for~12-18 hours for isolates with MPCs of 0.5 mg/L. In our study, 71% of isolates had MPC values to enrofloxacin �0.5 mg/L. Considering florfenicol, blood concentrations fall within the MSW, however, MPC values of �4 mg/L were seen for the majority of isolates tested and for these, drug concentration would exceed the MSW for~6 hours.
The MPC defines an in vitro measurement using high density bacterial populations that are representative of bacterial burdens present in acute infections. The MSW provide a drug concentration range where therapeutic drug concentrations falling and remaining in this range, based on approved dosage may allow for selective amplification of the least susceptible cells in the population. Dosing to achieve or exceed the MPC and hence the MSW prevents growth of bacterial cells with reduced susceptibility, however, dosing to exceed the MSW does not appear possible for all bug-drug combinations. In our study, enrofloxacin and ceftiofur had lower MIC and MPC values than did the other drugs. MPC principles may optimize therapy and impact resistance while providing valuable data for pharmacokinetic and pharmacodynamic modelling. Optimization of therapy while minimizing the potential for antimicrobial resistance are major principals for antimicrobial stewardship [49].